Broadcom

Headquarters

Palo Alto, California, USA

Website

www.broadcom.com

Employees

33,000 (FY2025 10-K)

Founded

1991

Listing

AVGO · NASDAQ

Market Cap

$1.82T

Revenue (TTM)

$75.46B

Price

USD383.22 (+1.34%)

Dependency

• Taiwan Semiconductor Manufacturing Company: Broadcom depends on TSMC to fabricate its custom silicon and networking chips.
• Apple: Apple and Broadcom extended their chip partnership through 2031 in a $30 billion deal announced July 2026, Apple's largest US manufacturing commitment to date, covering specialized chips for iPhones and Apple's Apple Intelligence server stack.

Supply Chain

• AI Data Centers: Broadcom supplies networking and semiconductor technologies used across advanced data centers.
• OpenAI: Broadcom co-developed OpenAI's first in-house AI inference chips, with the design phase complete and first data center deployment targeted for the second half of 2026.
• Alphabet (Google): Broadcom has expanded its custom AI chip design business to Google, alongside its existing TPU-adjacent partnership work.
• Anthropic: Broadcom has expanded its custom AI chip design business to Anthropic, confirming custom silicon as an industry-wide movement beyond its original OpenAI partnership.
• Meta: Meta is rolling out more than one gigawatt of custom silicon (MTIA) developed with Broadcom to complement its NVIDIA systems.

Development

• Co-Packaged Optics: Broadcom develops its own co-packaged optics technology for Ethernet switching, including its Tomahawk and Jericho platforms.

Enablement

• AI Accelerators: Broadcom develops semiconductor and networking technologies that support large-scale artificial intelligence computing infrastructure.

Competition / Other

• Marvell Technology: Broadcom and Marvell Technology compete directly for custom AI silicon (ASIC) design wins with hyperscale customers building their own accelerators.
• Cisco Systems: Broadcom competes with Cisco in networking silicon and switching products for AI data centers.
• Blackwell: Broadcom co-designs custom AI ASICs (including Google's TPU and Meta's MTIA) for hyperscalers seeking to reduce dependence on NVIDIA GPUs like Blackwell; a broader shift toward custom silicon for AI inference would directly grow Broadcom's addressable market at Blackwell's expense.
• NVIDIA: Broadcom participates in areas of AI infrastructure where custom silicon and networking technologies can compete with elements of NVIDIA's platform.